Privacy statistical method and system based on multi-party secure computation and trusted execution environment cooperation

Abstract

The application relates to the technical field of secure computation, and discloses a privacy statistical method and system based on multi-party secure computation and trusted execution environment cooperation, which comprises the following steps: performing fragmentation on threat intelligence statistical values collected by each security agency to obtain secret fragment values; calculating enclave local proof structure according to the enclave environment security parameters of a statistical node, and performing signature verification based on the enclave local proof structure to obtain a verification report; generating a session key based on the verification report, and performing encrypted transmission on the secret fragment values based on the session key to obtain encrypted secret fragments; reconstructing the encrypted secret fragments to obtain the threat intelligence statistical values, and performing multi-dimensional statistics on the threat intelligence statistical values to obtain threat intelligence statistical results; and through the secret fragmentation and secure enclave remote proof cooperation mechanism, the application realizes multiple security guarantees for threat intelligence privacy statistics.

Description

Technical Field

[0001] This invention relates to the field of secure computing technology, and in particular to a privacy statistics method and system based on the collaboration of multi-party secure computation and trusted execution environment. Background Technology

[0002] Sharing cyber threat intelligence is a key technological means to enhance cross-organizational security capabilities, enabling collaborative analysis and joint defense by aggregating threat data from multiple sources. However, existing threat intelligence statistics solutions generally adopt a centralized server model, requiring each organization to upload raw intelligence data to a unified platform for aggregation and calculation. This model poses a significant single point of trust risk; if the central server is compromised or the operator acts maliciously, it will lead to the complete leakage of sensitive intelligence data from all participating organizations.

[0003] While traditional privacy protection schemes based on encrypted computation can protect data confidentiality to a certain extent, they lack mechanisms to verify the trustworthiness of the computing environment. Malicious statistical nodes could forge the computing environment to steal encryption keys or intermediate computation results. Meanwhile, existing differential privacy schemes protect privacy through noise injection, but noise leads to a loss of statistical accuracy, affecting the accuracy of threat intelligence analysis and failing to meet the needs of critical infrastructure sectors for precise threat situation awareness. Summary of the Invention

[0004] This invention provides a privacy statistics method and system based on multi-party secure computation and trusted execution environment collaboration. This invention achieves multiple security guarantees for threat intelligence privacy statistics through secret sharding and secure enclave remote proof collaboration mechanism.

[0005] The first aspect of this invention provides a privacy statistics method based on the collaboration of multi-party secure computation and a trusted execution environment, the privacy statistics method based on the collaboration of multi-party secure computation and a trusted execution environment includes:

[0006] The statistical values ​​of threat intelligence collected by various security agencies are segmented to obtain secret segment values;

[0007] The enclave local proof structure is calculated based on the enclave environmental security parameters of the statistical nodes, and signature verification is performed based on the enclave local proof structure to obtain a verification report;

[0008] A session key is generated based on the verification report, and the secret fragment value is encrypted and transmitted based on the session key to obtain an encrypted secret fragment;

[0009] The encrypted secret fragment is reconstructed to obtain the threat intelligence statistics value, and the threat intelligence statistics value is statistically analyzed in multiple dimensions to obtain the threat intelligence statistics result.

[0010] In conjunction with the first aspect, in a first implementation of the first aspect of the present invention, the step of fragmenting the threat intelligence statistical values ​​collected by various security agencies to obtain secret fragment values ​​includes:

[0011] The reconstruction threshold parameter and the total number of collaborative statistical nodes are determined based on the total number of collaborative statistical participants, and the prime field modulus is selected as the finite field operation parameter.

[0012] Based on the reconstructed threshold parameters, a hardware random number generator is called to extract random mask coefficients. The threat intelligence statistics collected by each security agency are used as constant terms and the random mask coefficients are polynomially constructed under the prime field modulus to obtain the secret piecewise polynomial.

[0013] Based on the secret sharding polynomial, polynomial evaluation is performed on each node index value in the total number of collaborative statistical nodes to obtain the secret sharding value and sharding index number. Then, the secret sharding value is subjected to cyclic group exponent operation to obtain the sharding commitment value.

[0014] In conjunction with the first aspect, in a second implementation of the first aspect of the present invention, the step of calculating the enclave local proof structure based on the enclave environment security parameters of the statistical nodes, and performing signature verification based on the enclave local proof structure to obtain a verification report, includes:

[0015] The code pages in the threat statistics code binary file are loaded page by page into the enclave physical memory area, and a hash expansion and cumulative calculation is performed on each code page to obtain the enclave metric hash value;

[0016] The platform security version number and enclave attribute flag parameters are extracted from the processor platform of the statistical node to obtain the enclave environment security parameters;

[0017] Based on the enclave metric hash value and the enclave environment security parameters, a local proof data structure is encapsulated to obtain the enclave local proof structure;

[0018] The enclave local proof structure is encapsulated with a group signature to obtain a remote proof reference structure. The remote proof reference structure is then submitted to the proof verification service for signature verification to obtain a verification report.

[0019] In conjunction with the first aspect, in the third implementation of the first aspect of the present invention, the step of loading code pages from the threat statistics code binary file to the enclave physical memory region page by page, and performing hash expansion and cumulative calculation on each code page to obtain the enclave metric hash value, includes:

[0020] The enclave control structure allocation instruction is executed in the enclave physical memory region to obtain the enclave control structure, and the code pages of the threat statistics code binary file are loaded page by page into the enclave physical memory region based on the enclave control structure;

[0021] Each loaded code page in the enclave physical memory region is divided into blocks to obtain multiple code blocks, and a cumulative hash value update calculation is performed on each code block to obtain the cumulative hash value of the code page;

[0022] The accumulated hash value of the code page is locked to obtain the enclave metric hash value.

[0023] In conjunction with the first aspect, in the fourth implementation of the first aspect of the present invention, the step of encapsulating the enclave local proof structure with a group signature to obtain a remote proof reference structure, and submitting the remote proof reference structure to a proof verification service for signature verification to obtain a verification report, includes:

[0024] The enclave local proof structure is passed to the reference enclave for local proof validity verification, and the message data to be signed is extracted from the verified enclave local proof structure.

[0025] Based on the enhanced privacy group signature private key held by the reference enclave, the message data to be signed is encapsulated with a group signature to obtain a remote proof reference structure;

[0026] The remote proof reference structure is submitted to the proof verification service, which performs signature validity verification and platform security status checks on the group signature values ​​in the remote proof reference structure, and obtains a verification report.

[0027] In conjunction with the first aspect, in a fifth implementation of the first aspect of the present invention, the step of generating a session key based on the verification report and encrypting and transmitting the secret fragment value based on the session key to obtain an encrypted secret fragment includes:

[0028] The enclave metric hash value is extracted from the verification report and matched with the preset trusted code hash whitelist for verification. The platform security version number is extracted and compared with the preset minimum security baseline value for verification. The debug mode flag bit in the enclave attribute flag is extracted for verification of the closed status. After all verifications are passed, the statistical node is confirmed as a trusted node.

[0029] Each security agency and the enclave of the trusted node generate temporary elliptic curve key pairs and calculate a shared key, and perform key derivation processing on the shared key to obtain a session key;

[0030] Based on the session key, the secret fragment value, fragment index number and fragment commitment value are authenticated and encrypted, and a security agency identifier is attached as associated authentication data to obtain encrypted fragment ciphertext and authentication tag;

[0031] After receiving the encrypted fragment ciphertext and the authentication tag, the enclave decrypts the encrypted fragment ciphertext and performs tag verification on the authentication tag. After the tag verification is successful, the decrypted secret fragment value is stored in the enclave's isolated memory.

[0032] In conjunction with the first aspect, in the sixth implementation of the first aspect of the present invention, the reconstructing of the encrypted secret fragment to obtain the threat intelligence statistics value, and the multi-dimensional statistics of the threat intelligence statistics value to obtain the threat intelligence statistics result, includes:

[0033] Select the secret fragment value that reaches the reconstruction threshold parameter and the corresponding fragment index number from the enclave isolated memory, and perform reconstruction on the secret fragment value according to the fragment index number to obtain the reconstructed threat intelligence statistics value;

[0034] The reconstructed threat intelligence statistics are subjected to cyclic group exponent calculation and consistency verification with the fragmented commitment values. After the consistency verification is passed, the reconstruction is repeated for each security agency to obtain the threat intelligence statistics of each security agency.

[0035] The total number of attacks is obtained by summing the threat intelligence statistics of each security agency.

[0036] The statistical period is divided into time slots and the corresponding time slot count is incremented according to the attack timestamp to obtain a time distribution statistical array. A geographic location mapping query is performed on the attack source address and the corresponding region count is incremented to obtain a geographic distribution statistical mapping table.

[0037] Threat intelligence statistics are generated based on the total attack statistics, the time distribution statistics array, and the geographical distribution statistics mapping table.

[0038] In conjunction with the first aspect, in the seventh implementation of the first aspect of the present invention, the step of performing a cyclic group exponent calculation on the reconstructed threat intelligence statistics and verifying the consistency between the reconstructed threat intelligence statistics and the fragmented commitment values, and then repeatedly performing the reconstruction for each security agency after the consistency verification is passed, to obtain the threat intelligence statistics for each security agency, includes:

[0039] Based on preset cyclic group generator and group order parameters, cyclic group exponent operation is performed on the reconstructed threat intelligence statistics to obtain the reconstructed commitment value;

[0040] Each of the aforementioned piecewise commitment values ​​is subjected to exponentiation to obtain multiple exponentiation results, and modulo group order multiplication is performed on each exponentiation result to obtain the calculated commitment value;

[0041] The consistency between the reconstructed commitment value and the calculated commitment value is verified. After the consistency verification is passed, the correctness of the reconstruction is confirmed.

[0042] Based on the organization fragment receiving table maintained in the enclave isolated memory, the secret fragment value and fragment index number corresponding to each security organization are extracted one by one. The reconstruction is repeatedly performed for each security organization to obtain the threat intelligence statistics value of each security organization.

[0043] In conjunction with the first aspect, in the eighth implementation of the first aspect of the present invention, the privacy statistics method based on multi-party secure computation and trusted execution environment collaboration further includes:

[0044] A hardware-derived sealing key is generated based on the enclave signer identifier and security version number, and the threat intelligence statistics result is authenticated and encrypted based on the hardware-derived sealing key to obtain the ciphertext of the statistics result;

[0045] The integrity digest value is calculated based on the hardware-derived sealing key and the ciphertext of the statistical results, and the integrity digest value is digitally signed to obtain the digest signature value.

[0046] The encrypted statistical results, the integrity digest value, the digest signature value, and the sealing strategy parameters are used as the sealing statistical results, and the sealing statistical results are written to the persistent storage medium.

[0047] The system obtains and verifies an access token from the authorized authority. After the access token is verified, it decrypts the encrypted statistical results in the sealed statistical results and filters the access range to obtain the plaintext statistical results that the authorized authority can access.

[0048] A second aspect of the present invention provides a privacy statistics system based on the collaboration of multi-party secure computation and a trusted execution environment, the privacy statistics system based on the collaboration of multi-party secure computation and a trusted execution environment comprising:

[0049] The sharding module is used to shard the threat intelligence statistics collected by various security agencies to obtain secret shard values;

[0050] The signature verification module is used to calculate the enclave local proof structure based on the enclave environment security parameters of the statistical nodes, and to perform signature verification based on the enclave local proof structure to obtain a verification report.

[0051] An encrypted transmission module is used to generate a session key based on the verification report, and to encrypt and transmit the secret fragment value based on the session key to obtain an encrypted secret fragment;

[0052] The multi-dimensional statistics module is used to reconstruct the encrypted secret fragment, obtain the threat intelligence statistics value, and perform multi-dimensional statistics on the threat intelligence statistics value to obtain the threat intelligence statistics result.

[0053] Compared to existing technologies, this invention achieves multiple security safeguards for threat intelligence privacy statistics through a collaborative mechanism of secret sharding and secure enclave remote verification. Threshold polynomial sharding technology based on prime fields decomposes the threat statistics of each organization into multiple secret shards. Any shard with fewer than the threshold number cannot be used to deduce the original value algebraically, eliminating the risk of single-point data leakage and ensuring that the original intelligence data does not leave the local domain. MRENCLAVE metric values ​​are generated through page-by-page metric hash accumulation calculation of secure enclave code pages. Combined with enhanced privacy group signature encapsulation and a multi-layered remote verification mechanism using Intel IAS verification services, each organization can verify the code integrity, platform security version number, and debug mode status of the statistics node before sending shards, ensuring that the statistics environment has not been tampered with and is running on trusted hardware. The encrypted transmission of secret shards is combined with enclave-isolated memory storage, utilizing a hardware memory encryption engine to achieve ciphertext protection at the physical bus level. Even the operating system kernel or attackers with physical access rights cannot obtain the plaintext of the shards. The Lagrange interpolation reconstruction and cyclic group commitment value verification mechanism ensure the correctness of secret recovery. Addition aggregation, temporal distribution, and geographic mapping performed within the enclave occur without any intermediate data leakage. Encrypted storage based on hardware-derived sealing keys and a fine-grained authorization and desealing mechanism using access tokens achieves persistent protection of statistical results and on-demand access control, resolving the multiple conflicts between data sovereignty, environmental trust, and collaborative computing in cross-agency threat intelligence sharing. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] The structures, proportions, sizes, etc., shown in the accompanying drawings of this specification are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed in the specification, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0056] Figure 1 This is a flowchart illustrating the privacy statistics method based on the collaboration of multi-party secure computation and trusted execution environment provided in this embodiment of the invention.

[0057] Figure 2This is a schematic block diagram of the privacy statistics system based on the collaboration of multi-party secure computation and trusted execution environment provided in the embodiments of the present invention. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0060] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0061] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items, and all possible combinations, and includes such combinations. See also Figure 1 One embodiment of the privacy statistics method based on the collaboration of multi-party secure computation and trusted execution environment in this invention includes:

[0062] Step 110: Divide the threat intelligence statistics collected by various security agencies into segments to obtain secret segment values;

[0063] Specifically, the number of participants, k, in this round of collaborative statistical tasks is determined by each participating security organization, and a reconstruction threshold parameter t is set accordingly. The value of t satisfies 1 < t ≤ k. The original statistical value can only be recovered if at least t legitimate fragments participate in the reconstruction, forming the security foundation for threshold secret sharing. Simultaneously, a prime field modulus p that meets cryptographic strength requirements is selected, such as a 256-bit prime number, for use in all finite field modulo operations, enabling the construction of the polynomial and the calculation of fragments to have inverse computation capabilities. The organization uses the threat intelligence statistics value D_i to be protected as the constant term of the secret polynomial, and extracts t-1 random mask coefficients r1, r2, ..., r_{t-1} from it using a random number generator with a hardware entropy source, such as RDRAND, to construct a secret fragmentation polynomial of order t-1, P(x) = D_i + r1·x + r2·x. 2 + … + r_{t-1}·x^{t-1}, where mod represents the modulo operation, i.e., the remainder operation. All coefficients and the result of the operation are performed modulo p in the prime number field p, that is, the result of the calculation is the remainder of the prime number p, ensuring that the value is limited to the range of 0 to p-1. Even if the attacker intercepts part of the fragment, he will not be able to recover the value of D_i by algebraic reverse deduction. After constructing the polynomial, each institution, for the set of collaborative statistical node numbers {x1, x2, …, x_k}, substitutes each node index value x_j into P(x) to perform polynomial evaluation, calculating the corresponding secret fragment value S_j = P(x_j) mod p. That is, substituting x_j into the polynomial P(x) for calculation and performing a modulo p operation on the result yields the secret fragment value S_j. The secret fragment value S_j and its corresponding node index x_j are combined into an ordered pair, denoted as (S_j, j), where S_j represents the numerical content of the fragment, and j represents the node number corresponding to the fragment. By binding the fragment value with the index number, it ensures that the interpolation point position corresponding to each fragment can be correctly identified during the reconstruction process. Each institution uses the generator g of a preset cyclic group to calculate the commitment value C_j = g^{S_j} mod q for each secret fragment value on a multiplicative group of order q.

[0064] Step 120: Calculate the enclave local proof structure based on the enclave environmental security parameters of the statistical nodes, and perform signature verification based on the enclave local proof structure to obtain a verification report;

[0065] Specifically, a signed and encapsulated threat statistics code binary file is pre-deployed in the statistics node. This binary file is written page by page to the enclave's dedicated physical memory area, the EPC (Enclave Page Cache), during the loading process. The processor executes instruction set control operations such as ECREATE, EADD, and EEXTEND. When each page of code is written to the EPC, a hash expansion calculation based on SHA-256 is performed synchronously, adding the digest value of each 256-byte block to the cumulative hash register to form the overall metric hash value MRENCLAVE for the enclave. This metric hash value is unique and bound to the statistical logic and data structures running in the current enclave. The processor platform extracts the current platform's security version number SVN_platform and enclave attribute flags from the local security configuration register. These attribute flags contain parameters such as whether debugging is enabled in the enclave and whether it is in 64-bit mode. Together, these constitute the enclave environment security parameters to assess the trust level of the enclave instance at the hardware level. The statistics node calls the SGX instruction EREPORT to generate a local proof structure Report_local. It encapsulates the enclave metric hash value MRENCLAVE and enclave environment security parameters (including MRSIGNER, SVN_platform, and Attributes) into structured proof data. This data structure is then protected with a MAC signature using the local report key Report_Key derived from the processor root key, forming a hardware-authenticated enclave local proof structure. The statistics node then passes this local proof structure to the platform's built-in Quoting Enclave module. QuotingEnclave, authenticated by the Intel root signature key and possessing the processor's EPID group signature private key, hashes the metric value and environment parameter fields from the local proof structure as a message to be signed, generates a signature pair (R,s) using the EPID group signature algorithm, and encapsulates the signature along with the original metric information into a remote proof reference structure Quote. The statistics node submits the Quote to Intel's Attestation Service (IAS) verification service via HTTPS protocol. IAS performs group signature verification operations, including hash recalculation, signature correctness verification, platform status comparison and revocation list check. IAS generates a signature authentication verification report (AVR), which includes verification status fields, authentication timestamp, signed metric information, etc.

[0066] Step 130: Generate a session key based on the verification report, and encrypt and transmit the secret fragment value based on the session key to obtain the encrypted secret fragment;

[0067] Specifically, each security organization verifies the completeness and trustworthiness of the remote verification reports provided by the statistical nodes. This includes extracting the enclave metric hash value MRENCLAVE from the verification report and matching it item by item against a pre-defined trusted code hash whitelist to ensure that the currently running threat statistics code has not been tampered with and is indeed an approved version. The platform security version number SVN_platform is extracted and compared with a pre-defined minimum security baseline value to confirm that the hardware platform has patched critical side-channel vulnerabilities and has the ability to resist enclave leakage attacks. Simultaneously, the debug mode bit in the enclave attribute flag field Attributes is extracted to determine whether the debugging function is disabled, thus eliminating the potential leakage risk to the enclave memory caused by external debugging interfaces. Once all three verification conditions are met, the statistical node is determined to be a trusted node, and a secure communication channel is established with it. Each security organization generates a temporary elliptic curve key pair (sk_client, PK_client) using a high-strength random source and transmits the public key PK_client to the enclave via an unencrypted channel. The enclave synchronously generates its own temporary key pair (sk_enclave, PK_enclave) and returns the public key PK_enclave. Both organizations and the enclave execute the elliptic curve Diffie-Hellman key exchange algorithm using each other's public key and their local private key to calculate the shared key K_shared = ECDH(sk_client, PK_enclave). K_shared is then input into the key derivation function HKDF, and combined with a specified salt value and context information, a data session key K_session for authentication encryption is derived. The security organization uses K_session to perform authentication encryption on the locally generated secret fragment value S_j, the corresponding index number j, and the fragment commitment value C_j. The organization identifier institution_id is also included as additional authentication data AAD in the encryption process, resulting in the encrypted fragment ciphertext_j and the authentication tag tag_j. After receiving ciphertext_j and tag_j, the enclave uses K_session to decrypt ciphertext_j and performs authentication tag verification on tag_j. If tag verification fails, the fragment is immediately discarded to prevent inconsistency data injection; if verification passes, the decrypted (S_j, j, C_j) structured data is stored in the enclave's dedicated isolated memory area.

[0068] Step 140: Reconstruct the encrypted secret fragment to obtain threat intelligence statistics, and perform multi-dimensional statistics on the threat intelligence statistics to obtain threat intelligence statistics results.

[0069] Specifically, a fragment receiving table is maintained in the enclave-isolated memory within the statistical node. This table records the secret fragment values ​​uploaded by each security organization and authenticated after decryption, along with their corresponding fragment index numbers. When the number of valid fragments for a particular organization reaches the set reconstruction threshold parameter t, any t fragment values ​​S_j and their indices j are selected from the organization's fragment receiving records. Based on the Lagrange interpolation formula in Shamir threshold secret sharing, interpolation coefficients are calculated under a prime modulus p, and polynomial reconstruction is performed to recover the original threat intelligence statistics value D_i, which represents the attack event count detected by the organization within the statistical period. The enclave calculates a commitment value for the reconstructed D_i based on the cyclic group exponent operation rules. A weighted product consistency check is then performed between this commitment value and the commitment values ​​C_j corresponding to the t fragments involved in the reconstruction. This check verifies whether the commitment value is equal to the product of C_j^{L_j}. If the verification passes, it indicates that the fragments have not been tampered with and the reconstruction is correct. After successful verification, the enclave repeats the above reconstruction and verification process for all security agencies in sequence, restoring the threat intelligence statistics of all participants to form a statistical sequence {D_1, D_2, ..., D_k}. After obtaining the original statistical values ​​of all agencies, the enclave performs a total aggregation operation, summing all D_i to obtain the total attack statistics, which serves as the core indicator reflecting the overall cross-domain attack situation. Time distribution statistical analysis is performed on the attack timestamp sequences reported by each agency, dividing the statistical period into several fixed-width time slots (e.g., 48 slots in 30-minute units). The time slot index for each timestamp is calculated using modulo and integer division operations, and the corresponding slot counter is incremented to form an attack time distribution array. A pre-loaded geographic location database (e.g., GeoIP2 lightweight embedded version) is used to perform integer mapping and binary search on each attack source IP address to determine its country or region code, and the corresponding counter in the geographic distribution mapping table CountryCount[country_code] is incremented to form an attack geographic distribution dataset. Enclave encapsulates the total attack statistics, the time-series statistics of the TimeSlot array, and the geographical dimension statistics of CountryCount into a unified structure format, which is then output as the threat intelligence statistics result.

[0070] In one specific embodiment, the process of performing step 110 may specifically include the following steps:

[0071] The reconstruction threshold parameter and the total number of collaborative statistical nodes are determined based on the total number of collaborative statistical participants, and the prime field modulus is selected as the finite field operation parameter.

[0072] Based on the reconstructed threshold parameters, the hardware random number generator is called to extract the random mask coefficients. The threat intelligence statistics collected by various security agencies are used as constant terms and the random mask coefficients are used to construct a polynomial under the prime field modulus to obtain the secret fragmentation polynomial.

[0073] Based on the secret sharding polynomial, polynomial evaluation is performed on the index value of each node in the total number of collaborative statistical nodes to obtain the secret sharding value and sharding index number. Then, the cyclic group exponent operation is performed on the secret sharding value to obtain the sharding commitment value.

[0074] Specifically, during the system initialization phase, the platform coordinates and determines the total number of participants, k, in this round of collaborative statistical tasks. Based on the value of k, a secret reconstruction threshold parameter t is selected, where t satisfies 1 < t ≤ k. This ensures that the reconstruction threshold achieves both fault tolerance and prevents any number of fewer than t parties from jointly deriving the original data, thus providing resistance to collusion attacks. A prime field modulus p with sufficient cryptographic security is selected as the basis for finite field operations; for example, p should be greater than 2. 256 The 256-bit prime number possesses fast modulo arithmetic capabilities to support high-performance computing. After parameter selection, each security agency reads its locally collected threat intelligence statistics value D_i as a secret constant term. To prevent this value from being reverse-engineered during fragment construction, t-1 high-entropy mask coefficients r1, r2 to r_{t-1} are introduced. Each coefficient is independently generated by a hardware random number generator (such as Intel RDRAND or AMD RDSEED) with a physical entropy source guarantee, ensuring unpredictability and statistical uniformity. A t-1 order polynomial P(x) = D_i + r1·x + r2·x is constructed under the prime field modulus p. 2 + … + r_{t-1}·x^{t-1}, where all coefficients and variables are confined to GF(p) and modulo operations are performed, making the constructed polynomial mathematically hidden from the original values. After completing the secret sharding polynomial P(x), based on the total number of cooperative statistical nodes k, the node index values ​​x_j ∈ {1, 2, …,k} are enumerated in sequence. x_j is substituted into the polynomial P(x) to calculate the corresponding secret sharding value S_j = P(x_j) mod p, which together with x_j forms a sharding pair (S_j, j), and each shard is packaged independently. Each organization, based on the known generator g and group order q of a cyclic group, uses the fast modular exponentiation algorithm to perform an exponential mapping operation on the secret shard value S_j, calculating its shard commitment value C_j = g^{S_j}mod q. This shard commitment value is publicly released, but due to the intractability of the discrete logarithm problem, even if an attacker obtains C_j, they cannot recover S_j, thus achieving a balance between the concealment and verifiability of the shard content. After each shard is constructed, it, along with its index j and commitment value C_j, forms a structured encrypted shard unit.

[0075] The process includes obtaining the secret shard value and shard commitment value, followed by a distributed cross-validation step: each security agency broadcasts its generated shard commitment value to all other participating agencies via a broadcast protocol, and receives shard commitment values ​​broadcast by other agencies, thus obtaining a global set of shard commitment values; each security agency randomly selects a target agency to be verified from the global set of shard commitment values ​​based on a preset verification agency selection rule, generates a commitment verification challenge message containing a random challenge number and a verification timestamp for the target agency, and sends it to the target agency via a peer-to-peer encrypted channel; after receiving the commitment verification challenge message, the target agency executes the cross-validation on its own secret shard value based on the random challenge number. The zero-knowledge proof protocol is processed to obtain a verification response message containing the commitment response value and the zero-knowledge proof. The zero-knowledge proof is used to prove that the target organization holds a secret shard value corresponding to the shard commitment value and has not forged it. The verification response message is then returned to the security organization that initiated the verification. The security organization that initiated the verification performs a verification algorithm on the zero-knowledge proof in the received verification response message to verify the consistency between the commitment response value and the shard commitment value published by the target organization. If the verification passes, the target organization is marked as a legitimate organization and recorded in the trusted organization whitelist. If the verification fails, the target organization is marked as an abnormal organization and an abnormal warning message is broadcast to all participating organizations, triggering the shard removal process for abnormal organizations.

[0076] In one specific embodiment, the process of performing step 120 may specifically include the following steps:

[0077] The code pages in the threat statistics code binary file are loaded page by page into the enclave physical memory area, and a hash expansion and cumulative calculation is performed on each code page to obtain the enclave metric hash value;

[0078] The platform security version number and enclave attribute flag parameters are extracted from the processor platform of the statistical node to obtain the enclave environment security parameters;

[0079] The encapsulation of local proof data structure is based on encapsulation of encapsulation metric hash value and encapsulation environment security parameters to obtain encapsulation local proof structure;

[0080] The enclave local proof structure is encapsulated with a group signature to obtain a remote proof reference structure. The remote proof reference structure is then submitted to the proof verification service for signature verification to obtain a verification report.

[0081] Specifically, on statistics nodes supporting Intel SGX extensions, the dedicated binary file for threat statistics tasks is loaded. This binary file is an executable image compiled and signed using the SGX SDK toolchain, containing the code pages and data segments of the threat statistics logic. After initializing the enclave instance using the `sgx_create_enclave` instruction, the processor allocates control structures and initializes the security context for the Enclave using the `ECREATE` instruction. Subsequently, the `EADD` instruction is executed to write each page of data in the code segment page by page into the EPC (Enclave Page Cache) physically isolated memory region. For each page of code loaded, the processor automatically calls the `EEXTEND` instruction to perform a SHA-256 hash extension operation on each 256-byte block of the current page, and the calculation results are sequentially accumulated into internal registers. After all the code is loaded, the metric hash value `MRENCLAVE` corresponding to the Enclave is formed. Simultaneously, the processor extracts the platform security version number SVN_platform from the local platform configuration area to identify the firmware update level, and reads the Enclave attribute control word Attributes. This field records status bits such as whether the Enclave is in debug mode, whether it is running in 64-bit mode, and whether it is marked as measurable. These fields, together with MRENCLAVE, constitute the security parameter set of the current enclave environment. The enclave generates a local proof data structure Report_local by calling the EREPORT instruction. This structure encapsulates MRENCLAVE, MRSIGNER (signer's identity digest), the SVN_platform platform version number, and the Attributes attribute flags into a structure, and uses the local report key Report_Key derived from the processor's internal root key for MAC encryption authentication to ensure that this structure cannot be forged by the host operating system or external attackers. To enable remote security agencies to verify the trustworthiness of the enclave environment, the local proof structure is encapsulated by the QuotingEnclave processing module built into the SGX platform. The Quoting Enclave uses the processor's exclusive EPID group signature private key to sign the digest value in the local proof structure, generating a group signature pair (R, s), and then packages the signature, digest, and current timestamp together to form the remote proof reference structure Quote.The Quote structure features replay resistance, anonymous group signing capability, and verifiable code metrics. The statistics node submits the Quote to the Intel Attestation Service (IAS) verification service via a TLS encrypted channel. The IAS server uses the public key of the EPD group to verify the signature's legitimacy and cross-checks the platform status fields, including whether the platform has been revoked, whether the SVN_platform meets the minimum security baseline requirements, and whether the enclave is in non-debugging mode. After successful verification, an AVR verification report is generated, which includes the fields isvEnclaveQuoteStatus, MRENCLAVE value, timestamp, and digital signature.

[0082] In one specific embodiment, the process of loading code pages from the threat statistics code binary file to the enclave physical memory region page by page, and performing hash expansion and cumulative calculation on each code page to obtain the enclave metric hash value, can specifically include the following steps:

[0083] The enclave control structure allocation instruction is executed in the enclave physical memory region to obtain the enclave control structure, and the code pages of the threat statistics code binary file are loaded into the enclave physical memory region page by page based on the enclave control structure.

[0084] Each loaded code page in the enclave physical memory region is divided into blocks to obtain multiple code blocks, and the cumulative hash value update calculation is performed on each code block to obtain the cumulative hash value of the code page;

[0085] The accumulated hash value of the code page is locked to obtain the enclave metric hash value.

[0086] Specifically, on a processor platform with Intel SGX extension enabled, the ECREATE instruction is called to initialize an Enclave instance. During the initialization phase, the processor allocates an Enclave Control Structure (SECS) for the current enclave instance. This structure includes the Enclave base address, Enclave size, page attribute permissions, measurement status flags, signature message digest, etc., forming the basic control template for the physical and security operational boundaries of the entire enclave. The EADD instruction loads all code pages contained in the signed and encapsulated threat statistics code binary file into the EPC (Enclave Page Cache) area page by page. The loading order of all code pages follows the page layout specified in the signing phase, and each page is accompanied by a page access permission flag, such as read-only or executable. This permission information is written to the page table to prevent unauthorized behavior during Enclave operation. While writing the code page, the processor automatically triggers the EEXTEND instruction to perform block hashing on each page of code data, dividing each page into several fixed 256-byte code blocks. The SHA-256 hash function is executed on each block, and the resulting digest value is input into the cumulative metric register. After all blocks have been processed, the cumulative hash value of the current code page is formed. After all code pages are loaded and hash expansion is complete, the cumulative hash value is input as the metric digest into the lock register. The EINIT instruction is executed to confirm the enclave creation, and simultaneously locks the metric value MRENCLAVE as the integrity identifier of the Enclave instance.

[0087] In one specific embodiment, the process of encapsulating the enclave local proof structure with a group signature to obtain a remote proof reference structure, and submitting the remote proof reference structure to the proof verification service for signature verification to obtain a verification report, may specifically include the following steps:

[0088] The enclave local proof structure is passed to the reference enclave for local proof validity verification, and the message data to be signed is extracted from the verified enclave local proof structure;

[0089] Based on the enhanced privacy group signature private key held by the enclave, the group signature encapsulation is performed on the message data to be signed to obtain the remote proof reference structure;

[0090] The remote proof reference structure is submitted to the proof verification service, which verifies the signature validity and checks the platform security status of the group signature value in the remote proof reference structure, and obtains a verification report.

[0091] Specifically, the `Report_local` data structure is constructed by calling the SGX instruction `EREPORT`. This structure encapsulates the current enclave's metric hash `MRENCLAVE`, signer hash `MRSIGNER`, security version number `SVN`, attribute flags `Attributes`, and a user-specified digest field `report_data`. The entire structure is protected by MAC encryption using a processor-derived local reporting key. The statistics node transmits the local proof structure via the platform kernel driver to the Quoting Enclave, a dedicated enclave deployed on the SGX platform. This enclave is signed by Intel and holds the EPD (Enhanced Privacy ID) group signing private key, granting it privileged functionality for verifying the local report structure and performing remote proof conversion. Within the enclave, the incoming `Report_local` structure undergoes MAC verification using a processor-derived key. This verifies that the report indeed originates from the current instance on the platform and that the structure has not been tampered with. Upon successful verification, the message data fields to be signed are extracted, including `MRENCLAVE`, `MRSIGNER`, `SVN`, `Attributes`, `report_data`, the current timestamp, and a one-time random number `nonce`, which together form the message to be signed, `msg_to_sign`. The reference enclave invokes the EPD group signature module. Based on its held EPD private key sk_EPID, it performs a SHA-256 hash on msg_to_sign and generates a signature value (R, s) using the group signature algorithm. R is the product of the curve base points, and s is the response value. Together, they constitute the group signature body, possessing privacy features such as untraceable signer identity and prevention of replay attacks. The reference enclave encapsulates msg_to_sign and the signature value into a Quote structure, i.e., a remote proof reference structure, and sends the Quote structure to the Intel-provided Attestation Service (IAS) verification service via a TLS encrypted channel. The IAS server uses the publicly available EPD group public key to verify the signature in the Quote, recalculates the message hash, and verifies whether the elliptic curve signature equation holds, ensuring that the group signature was indeed generated by a trusted processor. After successful verification, the IAS checks the platform security status field to determine whether the platform model is on the revoked list, whether the security version number SVN is higher than the preset baseline, whether the debug mode in the attribute field is disabled, and confirms that the nonce in the Quote has not been reused, ensuring that the overall security status meets the trusted execution requirements. Once verification is complete, IAS generates an AVR verification report with a digital signature, which includes verification result fields such as isvEnclaveQuoteStatus, enclave metric hash MRENCLAVE, timestamp, signer identity, and platform status.

[0092] In one specific embodiment, the process of performing step 130 may specifically include the following steps:

[0093] The enclave metric hash value is extracted from the verification report and matched with the preset trusted code hash whitelist for verification. The platform security version number is extracted and compared with the preset minimum security baseline value for verification. The debug mode flag bit in the enclave attribute flag is extracted for verification of the closed status. After all verifications pass, the statistical node is confirmed as a trusted node.

[0094] Each security agency and trusted node's enclave generates temporary elliptic curve key pairs and calculates a shared key, and performs key derivation processing on the shared key to obtain the session key;

[0095] Based on the session key, the secret fragment value, fragment index number and fragment commitment value are authenticated and encrypted, and a security agency identifier is attached as associated authentication data to obtain the encrypted fragment ciphertext and authentication tag.

[0096] After receiving the encrypted fragment ciphertext and the authentication tag, the enclave decrypts the encrypted fragment ciphertext and verifies the authentication tag. After the tag verification is successful, the decrypted secret fragment value is stored in the enclave's isolated memory.

[0097] Specifically, each security organization receives a Remote Verification Report (AVR) from Intel Attestation Service. This report includes key fields such as the enclave metric hash MRENCLAVE, the platform security version number SVN_platform, and enclave attribute flags. Organizations parse the MRENCLAVE field and compare it item by item with a locally maintained trusted code hash whitelist to ensure that the code loaded by the enclave instance running on the current statistical node is indeed a security-audited and tamper-proof version. Simultaneously, the SVN_platform field is compared with a preset minimum security baseline value to ensure that the microcode version of the enclave platform is sufficient to defend against side-channel attack vulnerabilities, including L1TF and Foreshadow. The specific bit in the Attributes field used to identify debug status is checked to ensure it is turned off, i.e., debug mode is disabled. Once all three verifications pass, the enclave operating environment is confirmed as trustworthy. Each security organization calls its local elliptic curve cryptography library to generate a temporary key pair (sk_client, PK_client), and sends the public key PK_client to the enclave instance of the statistics node. The enclave also internally generates its own temporary key pair (sk_enclave, PK_enclave) and returns PK_enclave. Both parties use the received public key of the other party and their local private key to perform the ECDH key exchange algorithm to calculate the shared key K_shared = ECDH(sk_client, PK_enclave). The shared key is then input into the key derivation function based on HKDF, combined with context information and an application salt value, to derive a unique session key K_session. Each security organization uses the derived K_session to perform authentication encryption operations in AES-256-GCM mode on the locally constructed secret fragment value S_j, the corresponding fragment index number j, and the fragment commitment value C_j. The security organization identifier institution_id is used as the associated authentication data AAD in the encryption process. The encrypted fragment ciphertext and authentication tag are output and sent to the enclave via a TLS channel. After receiving the encrypted fragment ciphertext and authentication tag inside the enclave, it calls the decryption module to decrypt the encrypted fragment ciphertext using the session key and verifies the authentication tag. If the authentication tag does not match, it indicates that the data has been tampered with or forged, and storage is immediately rejected. Only after the tag verification is successful, the enclave writes the decrypted and recovered (S_j, j, C_j) structured fragment data into a dedicated isolated memory area.

[0098] In one specific embodiment, the process of performing step 140 may specifically include the following steps:

[0099] Select the secret fragment value that reaches the reconstruction threshold parameter and the corresponding fragment index number from the enclave isolated memory, and perform reconstruction on the secret fragment value according to the fragment index number to obtain the reconstructed threat intelligence statistics value;

[0100] The reconstructed threat intelligence statistics are subjected to cyclic group exponent calculation and fragment commitment value consistency verification. After the consistency verification is passed, the reconstruction is repeated for each security agency to obtain the threat intelligence statistics of each security agency.

[0101] The total number of attacks is obtained by summing the threat intelligence statistics of each security agency.

[0102] The statistical period is divided into time slots and the corresponding time slot count is incremented according to the attack timestamp to obtain a time distribution statistical array. A geographic location mapping query is performed on the attack source address and the corresponding region count is incremented to obtain a geographic distribution statistical mapping table.

[0103] Threat intelligence statistics are generated based on total attack statistics, time distribution statistics, and geographic distribution statistics mapping table.

[0104] Specifically, a structured fragment receiving table is maintained within the enclave, which records the valid secret fragments uploaded by each security agency, including the secret fragment value S_j, the corresponding fragment index number j, and its commitment value C_j. When the number of valid fragments corresponding to a certain agency reaches the preset reconstruction threshold t, t fragments are randomly selected from them for secret value recovery. Before performing the secret reconstruction, the enclave, based on the Shamir threshold secret sharing principle, extracts the index set {j1, j2, ..., j_t} from t fragments, and calculates the weight coefficient L_k corresponding to each fragment according to the Lagrange interpolation formula under the finite field modulus p, where L_k represents the weight coefficient of the k-th fragment in the Lagrange interpolation. Then, it solves the original secret Di = ∑{k=1}^{t} L_k·S{j_k} mod p, where S{j_k} represents the secret fragment value with index number j_k. This formula means multiplying all fragment values ​​S{j_k} from k to t by the corresponding weight coefficient L_k and summing them, and performing a modulo p operation on the summation result. mod means modulo operation, i.e., the remainder operation, which restores the threat intelligence statistics value originally submitted by the organization. To verify the correctness of the reconstructed values, the enclave calculates its cyclic group commitment C_reconstruct = g^{D_i} mod q, where mod q represents performing a modulo operation on the group order q. It then performs a commitment consistency check against the commitment values ​​of each fragment involved in the reconstruction, i.e., verifies whether the equation C_reconstruct = ∏{k=1}^{t} C{j_k}^{L_k} mod q holds true, where C{j_k} represents the commitment value corresponding to the fragment with index j_k, C{j_k}^{L_k} represents performing an L_k-th power operation on the commitment value C{j_k}, and the ∏ symbol represents a multiplication operation, i.e., multiplying all C{j_k}^{L_k} together. If true, it indicates that the fragment has not been tampered with and the reconstruction result is reliable; if false, the use of the agency statistics is terminated. This process is repeated for all security agencies sequentially, resulting in a statistical value sequence {D1, D2,..., D_k} within the enclave. The enclave performs an additive aggregation operation on the statistical value sequence to obtain the total attack statistics, which reflects the overall network threat intensity across the domain.To obtain the dynamic distribution of attacks over time, the enclave evenly divides the current statistical period into a fixed number of time slots (e.g., one slot every 30 minutes, 48 ​​slots in total). For each institution's uploaded attack timestamp set, the slot index corresponding to each timestamp within a 24-hour period is calculated as `slot_idx = floor((timestamp mod 86400) / 1800)`, where `floor` represents the floor operation, rounding down to the largest integer not greater than this value, `timestamp` represents the timestamp value of the attack, and `timestamp mod 86400`... 86400 indicates that the timestamp is modulo 86400 seconds (24 hours) to obtain the offset of seconds within the day. This offset is then divided by 1800 (30 minutes) and rounded down to obtain the time slot index number slot_idx to which the timestamp belongs. Finally, the statistical array TimeSlot[slot_idx] is incremented by 1. TimeSlot represents a statistical array used to record the number of attacks in each time slot. TimeSlot[slot_idx] represents the attack count value of the corresponding time slot for the element with index number slot_idx in the array. Incrementing the attack count of the time slot by 1 for this element forms a time distribution statistical array. Meanwhile, the enclave internally loads a lightweight geolocation IP mapping database, such as the GeoIP2 data structure. After converting the attack source IP addresses uploaded by various organizations into 32-bit integers, it uses a constant-time binary search method to match their corresponding country or region codes (country_code) to eliminate time-series side-channel attacks. Upon successful matching, the corresponding count value in the statistical mapping table CountryCount[country_code] is incremented. CountryCount represents a statistical mapping table used to record the number of attacks from each country or region, and CountryCount[country_code] represents the attack count value corresponding to the country or region code (key in the mapping table). Incrementing the attack count value increments the attack count for that country or region by 1, thus constructing a geographic attack distribution result. The enclave organizes the total attack statistics, time distribution statistics array, and geographic distribution statistics mapping table into a structured statistical result object, which is then encapsulated as threat intelligence statistics and returned to authorized access parties.

[0105] The process involves performing a geographic location mapping query on the attack source address and incrementing the corresponding region count to obtain a geographic distribution statistical mapping table. This includes an unintentional query step protected by a time-series side-channel: the geographic location database pre-stored in enclave isolation memory is divided into fixed-size data blocks for storage, resulting in a geographic location data block array. Each data block contains an address prefix range and the corresponding region code. A query mask vector is generated for the attack source address, with the length of the query mask vector equal to the total number of blocks in the geographic location data block array. A constant-time comparison operation is performed on each data block in the geographic location data block array, performing a bit-by-bit XOR operation between the binary representation of the attack source address and the address prefix range of the current data block. The accumulated XOR results yield a matching distance value, and the mask flag bit at the corresponding position in the query mask vector is updated based on whether the matching distance value is zero. The constant-time comparison operation ensures that the match is eliminated regardless of whether it is successful or not. It consumes the same number of processor clock cycles; it performs a mask selection operation on the geographic location data block array based on the query mask vector, traverses all data blocks and uses the mask flag bit to process conditional move instructions, and accumulates the region codes of the successfully matched data blocks into the result register. The conditional move instruction avoids timing differences caused by branch prediction at the hardware level, and obtains the query result region code; after obtaining the query result region code, it performs a random fill operation on the cache line in the enclave isolated memory, generates a random cache access pattern by accessing the preset cache noise data array, covers the cache traces caused by geographic location query, and performs timing fill processing on the entire query operation. By inserting no-operation instructions, it extends the actual query time to the preset constant time limit, eliminates the timing difference characteristics caused by different address queries, and prevents attackers from inferring the statistical geographic location distribution pattern through cache side channel or timing side channel.

[0106] In one specific embodiment, the process of performing a cyclic group exponent calculation on the reconstructed threat intelligence statistics and verifying the consistency of the fragmented commitment values, and then repeating the reconstruction for each security agency after the consistency verification is passed, to obtain the threat intelligence statistics for each security agency, can specifically include the following steps:

[0107] Based on preset cyclic group generator and group order parameters, cyclic group exponent operation is performed on the reconstructed threat intelligence statistics to obtain the reconstructed commitment value;

[0108] Perform exponentiation on each piecewise commitment value to obtain multiple exponentiation results, and perform modular group order multiplication on each exponentiation result to obtain the calculated commitment value;

[0109] The consistency between the reconstructed commitment value and the calculated commitment value is verified. After the consistency verification is passed, the correctness of the reconstruction is confirmed.

[0110] Based on the organization fragment receiving table maintained in the enclave isolated memory, the secret fragment value and fragment index number corresponding to each security organization are extracted one by one. The reconstruction is repeatedly performed for each security organization to obtain the threat intelligence statistics of each security organization.

[0111] Specifically, within the enclave, the secret reconstruction of the original threat intelligence statistics is completed. Based on the Shamir threshold secret sharing mechanism, fragment groups satisfying the reconstruction threshold t are selected from the fragment receiving table uploaded by the current organization. Each fragment value S_j and its corresponding index j are extracted, the Lagrange interpolation coefficients L_j are calculated, and the reconstruction result D_i = ∑L_j·S_j mod p is obtained in the prime field p. To verify the correctness of the reconstruction result, based on the generator g of the preset cyclic group and the corresponding group order q, a fast modular exponentiation operation is first performed on the reconstructed value D_i to calculate the reconstruction commitment value C_reconstruct = g^{D_i} mod q, obtaining the group element representation of the reconstructed statistics. For each original fragment's commitment value C_j participating in this reconstruction, an exponential operation is performed, i.e., C_j^{L_j} mod q is calculated. All these exponentiation results are then multiplied modulo q to obtain the calculated commitment value C_check = ∏ C_j^{L_j} mod q. This value is the theoretically expected commitment result based on the original commitment value and interpolation weights. C_reconstruct and C_check are compared for identity. If they match (C_reconstruct ≡ C_check mod q), it indicates that the selected fragment group is indeed a legitimate split of the original statistical values, and the reconstruction process has not been interfered with or tampered with, thus confirming the correct reconstruction of the threat intelligence statistics. After this verification, the enclave system, based on the fragment receiving tables maintained for each security agency in its internal isolated memory, extracts the secret fragment set and corresponding index for each agency, repeating the above reconstruction steps and commitment consistency verification process to sequentially reconstruct the original statistical values ​​D1, D2, ..., D_k uploaded by each security agency.

[0112] In one specific embodiment, the privacy statistics method based on the collaboration of multi-party secure computation and trusted execution environment further includes the following steps:

[0113] A hardware-derived sealing key is generated based on the enclave signer identifier and security version number. The threat intelligence statistics results are then authenticated and encrypted using the hardware-derived sealing key to obtain the ciphertext of the statistics results.

[0114] The integrity digest value is calculated based on the hardware-derived sealing key and the ciphertext of the statistical results, and the integrity digest value is digitally signed to obtain the digest signature value;

[0115] The encrypted statistical results, integrity digest value, digest signature value, and sealing strategy parameters are used as the sealed statistical results, and the sealed statistical results are written to persistent storage media.

[0116] Obtain and verify the access token from the authorized agency. After the access token is verified, decrypt the encrypted statistical results in the sealed statistical results and filter the access range to obtain the plaintext statistical results that the authorized agency can access.

[0117] Specifically, within the enclave, the EGETKEY instruction is invoked to generate a sealed key bound to the current Enclave instance. The instruction input parameters include the KeyName field (set to SEAL_KEY), the KeyPolicy field (set to MRSIGNER), and the current enclave's signer identity digest MRSIGNER and security version number ISSSVN. Based on these parameters, the processor derives a unique and non-derivative sealed key, Seal_Key, from the hardware root key. This key can only be recovered and used within the same signer and the same or higher version of the enclave. The enclave uses the threat intelligence statistics result structure Result_data as the plaintext to be protected, and performs authentication encryption using Seal_Key combined with the AES-256-GCM encryption algorithm. The encrypted appended data is specified as a combination of the enclave metric hash and a timestamp, resulting in the encrypted statistical result ciphertext result_ciphertext and an authentication tag. The enclave performs HMAC-SHA256 calculations on the ciphertext of the statistical results and the encrypted appended data using the same sealing key, generating an integrity digest value (Digest). It then calls the ECDSA signature module within the Enclave to perform a digital signature operation on the digest value, obtaining a digest signature value (Sign_result) to ensure the ciphertext has not been tampered with. The enclave packages the ciphertext of the statistical results, the integrity digest value, the digest signature value, the encrypted appended data (AAD), the sealing policy (KeyPolicy), and the version information (ISVSVN) into a sealed statistical result (Sealed_result), and writes this structure to non-volatile persistent storage media such as a local NVMe hard drive or a remote encrypted storage server. Upon receiving an access token (Token) containing the organization's identity, access level, authorization fields, and signature from the querying party, the enclave verifies the token's signature validity using a pre-configured authorization center public key and checks whether the token has expired, whether the organization is allowed access, and whether the access level overrides the current request. After successful verification, the enclave calls EGETKEY to re-export the sealing key Seal_Key and performs AES-GCM decryption to obtain the plaintext statistical results. It calculates the HMAC digest value and compares it with the original digest to ensure that it has not been tampered with. Based on the access scope field described in the Token, it filters the statistical data in the plaintext and only returns a subset of fields that are allowed to be accessed, such as total attack volume, time distribution, and geographical distribution. It uses the query public key provided by the authorized agency to perform ECIES encryption to form a response data packet, which is returned to the querying party through the TLS secure channel.

[0118] The process involves several steps, including fragment integrity verification and fault-tolerant retransmission, after the enclave receives the encrypted fragment ciphertext and authentication tag, before decryption and tag verification. These steps include: performing transport layer integrity verification on the received encrypted fragment ciphertext, calculating the hash fingerprint value of the encrypted fragment ciphertext, comparing it with the original fingerprint value attached by the sending organization before transmission, detecting bit flips or data corruption during transmission, and marking the fragment as a corrupted fragment and recording it in the abnormal fragment log, triggering a retransmission request to the sending organization. After the transport layer integrity verification passes, performing mathematical validity verification on the decrypted secret fragment value, verifying whether the secret fragment value is within a preset prime number modulus range, and performing a pre-verification commitment equation calculation based on the fragment index number and fragment commitment value. The equation relationship is verified through cyclic group exponentiation. If the secret fragment value exceeds the valid range or the commitment equation verification fails, the fragment is marked as a tampered fragment and the sending organization identifier is recorded, generating an abnormal fragment report containing the anomaly type and detection timestamp. Based on the abnormal fragment... The system generates fragment logs and abnormal fragment reports, performs statistical analysis on the fragment anomaly rate of each security organization, calculates the abnormal fragment percentage for each organization, and adds the organization to the suspected malicious organization watchlist when the abnormal fragment percentage of an organization exceeds a preset anomaly threshold. It also suspends receiving subsequent fragments from that organization and sends an organization reputation assessment request to the collaborative statistical management node, triggering a security audit process for that organization. For fragments marked as corrupted, the enclave generates a fragment retransmission request message containing the fragment index number, corruption timestamp, and retransmission random number, and sends it to the corresponding security organization through an established session key encrypted channel. After receiving the retransmission request message, the security organization re-encapsulates the encrypted fragment based on the stored original secret fragment value and retransmits it. After receiving the retransmitted fragment, the enclave repeatedly performs integrity and legality checks. If multiple retransmissions fail or all of the organization's fragments are abnormal, all of the organization's fragments are removed from the enclave's isolated memory, and the organization is removed from the list of organizations participating in the statistics, ensuring that statistical calculations are performed only on legitimate and complete secret fragments.

[0119] The privacy statistics method based on the collaboration of multi-party secure computation and trusted execution environment in the embodiments of the present invention has been described above. The privacy statistics system based on the collaboration of multi-party secure computation and trusted execution environment in the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 2 One embodiment of the privacy statistics system based on multi-party secure computation and trusted execution environment collaboration in this invention includes:

[0120] The fragmentation module 210 is used to fragment the threat intelligence statistics collected by various security agencies to obtain secret fragment values.

[0121] The signature verification module 220 is used to calculate the enclave local proof structure based on the enclave environment security parameters of the statistical nodes, and to perform signature verification based on the enclave local proof structure to obtain a verification report.

[0122] The encrypted transmission module 230 is used to generate a session key based on the verification report, and to encrypt and transmit the secret fragment value based on the session key to obtain an encrypted secret fragment;

[0123] The multi-dimensional statistics module 240 is used to reconstruct the encrypted secret fragments, obtain threat intelligence statistics values, and perform multi-dimensional statistics on the threat intelligence statistics values ​​to obtain threat intelligence statistics results.

[0124] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0125] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0126] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A privacy statistics method based on the collaboration of multi-party secure computation and trusted execution environment, characterized in that, include: The statistical values ​​of threat intelligence collected by various security agencies are segmented to obtain secret segment values; Specifically, this includes: determining the reconstruction threshold parameter and the total number of collaborative statistical nodes based on the total number of collaborative statistical participants, and selecting the prime field modulus as the finite field operation parameter; based on the reconstruction threshold parameter, calling a hardware random number generator to extract random mask coefficients, using the threat intelligence statistics collected by each security agency as a constant term and the random mask coefficients under the prime field modulus to construct a polynomial, thereby obtaining a secret sharding polynomial; based on the secret sharding polynomial, performing polynomial evaluation calculations on each node index value in the total number of collaborative statistical nodes to obtain a secret sharding value and a sharding index number, and performing cyclic group exponent operation on the secret sharding value to obtain a sharding commitment value; The enclave local proof structure is calculated based on the enclave environmental security parameters of the statistical nodes, and signature verification is performed based on the enclave local proof structure to obtain a verification report; A session key is generated based on the verification report, and the secret fragment value is encrypted and transmitted based on the session key to obtain an encrypted secret fragment; The encrypted secret fragment is reconstructed to obtain the threat intelligence statistics value, and the threat intelligence statistics value is statistically analyzed in multiple dimensions to obtain the threat intelligence statistics result.

2. The privacy statistics method based on multi-party secure computation and trusted execution environment collaboration as described in claim 1, characterized in that, The process involves calculating the enclave local proof structure based on the enclave environmental security parameters of the statistical nodes, and performing signature verification based on the enclave local proof structure to obtain a verification report, including: The code pages in the threat statistics code binary file are loaded page by page into the enclave physical memory area, and a hash expansion and cumulative calculation is performed on each code page to obtain the enclave metric hash value; The platform security version number and enclave attribute flag parameters are extracted from the processor platform of the statistical node to obtain the enclave environment security parameters; Based on the enclave metric hash value and the enclave environment security parameters, a local proof data structure is encapsulated to obtain the enclave local proof structure; The enclave local proof structure is encapsulated with a group signature to obtain a remote proof reference structure. The remote proof reference structure is then submitted to the proof verification service for signature verification to obtain a verification report.

3. The privacy statistics method based on multi-party secure computation and trusted execution environment collaboration as described in claim 2, characterized in that, The process of loading code pages from the threat statistics code binary file into the enclave's physical memory region page by page, and performing hash expansion and cumulative calculation on each code page to obtain the enclave metric hash value, includes: The enclave control structure allocation instruction is executed in the enclave physical memory region to obtain the enclave control structure, and the code pages of the threat statistics code binary file are loaded page by page into the enclave physical memory region based on the enclave control structure; Each loaded code page in the enclave physical memory region is divided into blocks to obtain multiple code blocks, and a cumulative hash value update calculation is performed on each code block to obtain the cumulative hash value of the code page; The accumulated hash value of the code page is locked to obtain the enclave metric hash value.

4. The privacy statistics method based on multi-party secure computation and trusted execution environment collaboration as described in claim 3, characterized in that, The enclave-local proof structure is encapsulated with a group signature to obtain a remote proof reference structure. This remote proof reference structure is then submitted to a proof verification service for signature verification, resulting in a verification report, including: The enclave local proof structure is passed to the reference enclave for local proof validity verification, and the message data to be signed is extracted from the verified enclave local proof structure. Based on the enhanced privacy group signature private key held by the reference enclave, the message data to be signed is encapsulated with a group signature to obtain a remote proof reference structure; The remote proof reference structure is submitted to the proof verification service, which performs signature validity verification and platform security status checks on the group signature values ​​in the remote proof reference structure, and obtains a verification report.

5. The privacy statistics method based on multi-party secure computation and trusted execution environment collaboration as described in claim 1, characterized in that, The step of generating a session key based on the verification report and encrypting and transmitting the secret fragment value based on the session key to obtain an encrypted secret fragment includes: The enclave metric hash value is extracted from the verification report and matched with the preset trusted code hash whitelist for verification. The platform security version number is extracted and compared with the preset minimum security baseline value for verification. The debug mode flag bit in the enclave attribute flag is extracted for verification of the closed status. After all verifications are passed, the statistical node is confirmed as a trusted node. Each security agency and the enclave of the trusted node generate temporary elliptic curve key pairs and calculate a shared key, and perform key derivation processing on the shared key to obtain a session key; Based on the session key, the secret fragment value, fragment index number and fragment commitment value are authenticated and encrypted, and a security agency identifier is attached as associated authentication data to obtain encrypted fragment ciphertext and authentication tag; After receiving the encrypted fragment ciphertext and the authentication tag, the enclave decrypts the encrypted fragment ciphertext and performs tag verification on the authentication tag. After the tag verification is successful, the decrypted secret fragment value is stored in the enclave's isolated memory.

6. The privacy statistics method based on multi-party secure computation and trusted execution environment collaboration as described in claim 5, characterized in that, The process involves reconstructing the encrypted secret fragment to obtain the threat intelligence statistics value, and then performing multi-dimensional statistics on the threat intelligence statistics value to obtain the threat intelligence statistics result, including: Select the secret fragment value that reaches the reconstruction threshold parameter and the corresponding fragment index number from the enclave isolated memory, and perform reconstruction on the secret fragment value according to the fragment index number to obtain the reconstructed threat intelligence statistics value; The reconstructed threat intelligence statistics are subjected to cyclic group exponent calculation and consistency verification with the fragmented commitment values. After the consistency verification is passed, the reconstruction is repeated for each security agency to obtain the threat intelligence statistics of each security agency. The total number of attacks is obtained by summing the threat intelligence statistics of each security agency. The statistical period is divided into time slots and the corresponding time slot count is incremented according to the attack timestamp to obtain a time distribution statistical array. A geographic location mapping query is performed on the attack source address and the corresponding region count is incremented to obtain a geographic distribution statistical mapping table. Threat intelligence statistics are generated based on the total attack statistics, the time distribution statistics array, and the geographical distribution statistics mapping table.

7. The privacy statistics method based on multi-party secure computation and trusted execution environment collaboration as described in claim 6, characterized in that, The reconstructed threat intelligence statistics are subjected to cyclic group exponentiation and consistency verification with the fragmented commitment values. After the consistency verification is passed, the reconstruction is repeated for each security agency to obtain the threat intelligence statistics for each security agency, including: Based on preset cyclic group generator and group order parameters, cyclic group exponent operation is performed on the reconstructed threat intelligence statistics to obtain the reconstructed commitment value; Each of the aforementioned piecewise commitment values ​​is subjected to exponentiation to obtain multiple exponentiation results, and modulo group order multiplication is performed on each exponentiation result to obtain the calculated commitment value; The consistency between the reconstructed commitment value and the calculated commitment value is verified. After the consistency verification is passed, the correctness of the reconstruction is confirmed. Based on the organization fragment receiving table maintained in the enclave isolated memory, the secret fragment value and fragment index number corresponding to each security organization are extracted one by one. The reconstruction is repeatedly performed for each security organization to obtain the threat intelligence statistics value of each security organization.

8. The privacy statistics method based on multi-party secure computation and trusted execution environment collaboration according to claim 1, characterized in that, The privacy statistics method based on the collaboration of multi-party secure computation and trusted execution environment also includes: A hardware-derived sealing key is generated based on the enclave signer identifier and security version number, and the threat intelligence statistics result is authenticated and encrypted based on the hardware-derived sealing key to obtain the ciphertext of the statistics result; The integrity digest value is calculated based on the hardware-derived sealing key and the ciphertext of the statistical results, and the integrity digest value is digitally signed to obtain the digest signature value. The encrypted statistical results, the integrity digest value, the digest signature value, and the sealing strategy parameters are used as the sealing statistical results, and the sealing statistical results are written to the persistent storage medium. The system obtains and verifies an access token from the authorized authority. After the access token is verified, it decrypts the encrypted statistical results in the sealed statistical results and filters the access range to obtain the plaintext statistical results that the authorized authority can access.

9. A privacy statistics system based on the collaboration of multi-party secure computation and trusted execution environment, characterized in that, A privacy statistics method for performing any one of claims 1-8 based on multi-party secure computation and trusted execution environment collaboration includes: The sharding module is used to shard the threat intelligence statistics collected by various security agencies to obtain secret shard values; The signature verification module is used to calculate the enclave local proof structure based on the enclave environment security parameters of the statistical nodes, and to perform signature verification based on the enclave local proof structure to obtain a verification report. An encrypted transmission module is used to generate a session key based on the verification report, and to encrypt and transmit the secret fragment value based on the session key to obtain an encrypted secret fragment; The multi-dimensional statistics module is used to reconstruct the encrypted secret fragment, obtain the threat intelligence statistics value, and perform multi-dimensional statistics on the threat intelligence statistics value to obtain the threat intelligence statistics result.

Images